Current testing is done exclusivly with pitzer.dat, but it was observed that other databases could produce unexpected results. This was partly due to commission of pH balance on any of block with exact speciation and observed for llnl.dat and wateq4f database. This would typically result in unexpected pH changes for exact speciated block ( adding acid increases pH rather then drops it).
The PR on ph Balance should solve the problem, but needs explicit benchmarks.